- 75
- 1 473
Bill
เข้าร่วมเมื่อ 27 ส.ค. 2013
Week 7: Friendly Dispersion
In this final episode we enhance our model with a richer and more textured of our data and hypotheses. We transform the Poisson count model to accommodate over-dispersed data like typhoons and innovation. We use the Gamma distribution to generate Poisson rates. We also return to the tool count model from Oceanic cultures to investigate the impact of changing our observational model from Poisson to gamma-Poisson. The Gamma distribution is one of two fundamental models in probability and statistics, along with the Exponential model. The Gamma is indeed one of the exponential family of models if there were a hierarchy. The Gamma can be transformed into the Chi-squared and Normal (Gaussian) models without much ado. Base R plots are deployed to visualize our results and help us tell our story. Indications of next steps will open into new conversations about communities of interest, clusters, distance metric for contact. Much to learn! Enjoy!
มุมมอง: 0
วีดีโอ
Week 6: Sharing Food - Social Network Analysis
มุมมอง 212 ชั่วโมงที่ผ่านมา
As we enter the last phase of this course we begin to think about topics for a final project to be documented in a paper. Here is the beginning of such an example which grounds itself in existing research explicated by McElreath in chapter 14, section 3 on food sharing networks in Nicaragua. We build, analyze, and visualize houses and dyadic gift exchanges using multivariate normal priors on ge...
Week 5 - Video 4: Eagle Pirates and Gender Walkabouts
มุมมอง 119 ชั่วโมงที่ผ่านมา
We walk through the Week 5 Skill Builders with more binomial and poisson practice. Imagine bald eagles scavenging off other bald eagles for salmon. Also imagine gender discrimination in awarding grants. We tackle these issues using Stan, ulam, rethinking in base R, to tell our stories.
Week 5 - Video 3: Poisson Oceanic Tools
มุมมอง 119 ชั่วโมงที่ผ่านมา
We meet up with toolmakers in Oceania around the time of European contact. Does high contact and large populations matter to tool counts? We build a simple technology production model with some memory of past innovations and generate probable counts consistent with data and theory. We also delve into the Stan Probabilistic Programming environment, become familiar with basic model building and s...
Week 5 - Video 2: Project Notes
มุมมอง 6วันที่ผ่านมา
In this episode we talk about the final project for the course. A project workflow is proposed. Included are an introduction with the key questions posed for analysis, exploratory data analysis, causal inference, and concluding discussion. We will use a bib.tex file to encode information about the information we use in the report.
Week 5 - Video 1: Binomial Social Primates
มุมมอง 18วันที่ผ่านมา
In this episode we compare and contrast actual to predicted probabilities of social success for primates and a stop by to examine mayfly occupancy in bentic environments. We deploy an actor-treatment binomial regression with treatment impacts to understand difference, similarities, and indifference among respondents.
Week 4 - Video 3: Tidy Patchwork
มุมมอง 1514 วันที่ผ่านมา
In this episode we walk through the modeling of a burning question: what about the Seychelles? Does terrain ruggedness impact economic development? And how do clear outliers like Austria and The Seychelles influence the impact? We build a straight=forward probabilistic regression model of gdp (real-per capita) on ruggedness (terrain ruggedness index) across two regions: Africa and the rest of t...
Week 4- Video 2: Just Coffee
มุมมอง 614 วันที่ผ่านมา
In this episode we rework the spreadsheet model of video 1 using a customer-coffee-shift time (am-pm) example. We deploy our pose-hypothesize-sample-count-infer analytical process and generate visuals to answer the question: how do different shifts relate to customer who only want coffee? The sampling is simple, the counting is binomial (again), the infer is simulating the counting and producin...
Week 4 - Video 1: Seriously Dazed and Confused
มุมมอง 2421 วันที่ผ่านมา
A very simple tale using just a spreadsheet in this episode. We ask ai binomial style question about the number of serious incidents on the block where I live during the past month. We arrive at a hypothesized proportion of serious incidents out of total incidents using exactly the same analysis as in Week 1 with Earth and Mars. This time we get a little clever and see that the proportion of se...
Week 3 - Video 4: Forks in the Road While Worlds Collide
มุมมอง 621 วันที่ผ่านมา
In this last episode for Week 3 we use jus t wo rules to discover reasonable and minimal sets of regressors, explanatory variables, to support our search for how exposures can influence outcomes. Of course we are talking about DAGs, directed (arrows) acyclic (no feedback) graphs (vertices and edges). The rules are this simple. 1. Lose the arrows and trace all paths from exposure (X) to outcome ...
Week 3 - Video 3: WAICing Up to Inference
มุมมอง 921 วันที่ผ่านมา
In this episode we encounter information theoretic measures of uncertaintly including log(odds ratios), Kullback-Leibler divergences, and the Watanabe-Akaike Information Criteria (WAIC) for model selection. We trade off Pareto-Smoothed Importance Sampling (PSIS) with leave-one-out (LOO) validation, uncertainty trading off with volatility. At least a soupcon of an introduction awaits us.
Week 3 - Video 2: Rushing to the Logical
มุมมอง 421 วันที่ผ่านมา
We discover new insights through a binary logic and its incredibly simple algebra. We work through key conditional inferences and discover Bayes Rule in the middle of the haystack. A step-by-step logic flow guides us to making inferences consistent with data.
Week 3 - Video 1: Cats Sum It Up
มุมมอง 1521 วันที่ผ่านมา
In this episode we revisit colliders, forks, chains and Bayes inference through the Algebra of Boole and the Directed Acyclic Graph probabilistic approach of Judea Pearl. We walk through a 9th grade elementary algebra approach to the analysis of complex logical statements, to each of which we assign probability values. From there we infer from a scientifically causal model what the statistical ...
Week 2 - Video 4: Walking through the assignment
มุมมอง 1728 วันที่ผ่านมา
We cover how to find flaws in our priors and fix them; mediate otherwise confounding explanations; simulate the uncertainty of influences; transform variables so that a model more consistently depicts data. Always we conform our mind (models) with reality (data).
Week 2 / Video 3: Causal Cats Are Prowling
มุมมอง 4หลายเดือนก่อน
We begin our principled approach to model construction with Directed Acyclic Graphs, their generative modeling, and how to deal with forks, chains (pipes), colliders, and descendants. We play with our synthetic models a la Gilder and Schragge (waste transistors and simulations!). Behind the scenes is George Bools''s binary algebra to keep up honest about inferences, and Charles Peirce's retrodu...
Week 2 / Video 2: 2 faces of inference - Frequentists and Bayesians
มุมมอง 5หลายเดือนก่อน
Week 2 / Video 2: 2 faces of inference - Frequentists and Bayesians
Week 1 / LIVE SESSION 1: globe tossing for fun and profit
มุมมอง 12หลายเดือนก่อน
Week 1 / LIVE SESSION 1: globe tossing for fun and profit
Week 1 / Video 3: Sample until we learn something!
มุมมอง 31หลายเดือนก่อน
Week 1 / Video 3: Sample until we learn something!
Week 1 / Video 2: Counting (on Bayes) with R
มุมมอง 25หลายเดือนก่อน
Week 1 / Video 2: Counting (on Bayes) with R
WELCOME to Probabilistic Reasoning (aka Bayesian Data Analytics)
มุมมอง 23หลายเดือนก่อน
WELCOME to Probabilistic Reasoning (aka Bayesian Data Analytics)
LIVE SESSION 7: simulating optimal design of a self-insurance fund
มุมมอง 19หลายเดือนก่อน
LIVE SESSION 7: simulating optimal design of a self-insurance fund
LIVE SESSION 6: high impact-high uncertainty simulation of possible outcomes
มุมมอง 17หลายเดือนก่อน
LIVE SESSION 6: high impact-high uncertainty simulation of possible outcomes
LIVE SESSION 5: building a simple waiting time generative AI
มุมมอง 5หลายเดือนก่อน
LIVE SESSION 5: building a simple waiting time generative AI
LIVE SESSION 4: end-to-end multiperiod supply chain planning
มุมมอง 52 หลายเดือนก่อน
LIVE SESSION 4: end-to-end multiperiod supply chain planning
LIVE SESSION 3: a simple 2x2 decision model
มุมมอง 202 หลายเดือนก่อน
LIVE SESSION 3: a simple 2x2 decision model
LIVE SESSION 2: Adventures in Solar Leasing
มุมมอง 542 หลายเดือนก่อน
LIVE SESSION 2: Adventures in Solar Leasing
LIVE SESSION 1. Spreadsheet Basics #1: Associative Tables and Grids
มุมมอง 182 หลายเดือนก่อน
LIVE SESSION 1. Spreadsheet Basics #1: Associative Tables and Grids